i.

The traditional method for the design against buckling of cylindrical shells subjected to a
general load case (causing meridional, circumferential and/or shear stresses) requires the
calculation of the elastic critical shell buckling stresses and of appropriate shell buckling
knockdown factors for the individual stress components. Appropriate “hand formulae”
from the literature - or the Annex of the code itself - can be used for the calculation of
these quantities. The interaction between the different stress components is dealt with
through an interaction formulation.
ii. The increase of availability and expertise in the use of advanced numerical analysis
techniques and software packages was acknowledged in the code by the introduction of
the so-called “MNA/LBA Method” or “Overall Method”. This approach is found in
section 8.6 of EN 1993-1-6 - “Design by global numerical analysis using MNA and LBA
analyses”. The acronyms used represent a Materially Nonlinear Analysis, used to
determine the plastic limit load (ignoring instability), and a Linear Buckling Analysis,
used to calculate the linear elastic bifurcation load of the shell (ignoring imperfections
and plasticity). The name “Overall Method” stems from the use of an “overall”
slenderness, valid for the whole structure and the studied loading condition, and
calculated on the basis of the global MNA and LBA analyses. The basic principles of the
method are summarized in Fig. 3 and are described in section 2.2.
iii. Finally, it shall be mentioned that EN1993-1-6 also contains an additional, even more
advanced method, which makes use of analysis techniques that are Geometrically and
Materially Nonlinear and account for Imperfections – GMNIA. This represents a full
elasto-plastic finite displacement numerical (FEM) analysis with appropriate geometric
and structural imperfection definitions. The method is used in later sections of this paper
as a research tool. For the design of typical cylindrical shells against buckling, the code
recommends great care and requires validation effort by the structural engineer who
wishes to employ it. Contrary to the “traditional” and the “Overall – MNA/LBA”
approaches, it is not generally seen as a design tool, but rather as a tool to verify existing
structures in certain cases and applications, as well as a method suitable for code-making
and applied research.

Figure 3: The “Overall” or “MNA/LBA Method” in EN 1993-1-6

3
 Figure 4: Range of buckling reduction factors =ov (a); influence of the steel grade on the reduction factor x for
meridional compression.

Table 1: Coefficients for Eq. 2a-2c for basic buckling cases

Buckling case  ov,0   

meridional 0.21 Eq. 3 +

compression x Table 2
circumferential 0.4 Table 2 0.60 1.00
compression 
shear x 0.4 Table 2
1. Can be set to 0.3 for cylinders in bending when certain conditions are met.

0.62
for meridional compression: x  (3a)
1  1.91   w k / t 
1.44

1 r
with w k / t   (3b)
Q t

Table 2: Coefficients dependent on the “Quality Classes”

Buckling case Class A Class B Class C

meridional Q=40 Q=25 Q=16

compression x
circumferential =0.75 =0.65 =0.50
compression 
shear x =0.75 =0.65 =0.50

5
3. High-Strength Steel Cylindrical Hollow Sections – “Slender” Cylinders in Compression?
3.1 General considerations
The classical fields of application of the shell buckling rules of EN 1993-1-6 were already
pointed out in section 1, Fig. 1 – i.e. tanks, silos, large-diameter masts and penstocks, etc.
However, a new field of application of the shell buckling rules may become of relevance: the
determination of the local buckling strength of circular hollow sections (CHS) made of high-
strength (HS) steel. In Europe, CHS are produced as hot-finished or cold-formed welded sections
in accordance with the product standards EN 10210 and EN 10219 (CEN 2006a, 2006b),
respectively. With few exceptions, commonly used diameters do not exceed 600mm in the case
of hot-finished and 1000mm in the case of welded sections produced to these standards; the vast
majority of employed sections are far smaller, and typical diameter-to-thickness D/t ratios stay
well below 100. CHS made of mild steel grades are usually designed in accordance with design
standards such as the general part 1-1 of EN 1993 (CEN 2005), the AISC (2010) specification
for steel buildings, CIDECT (2010) design guidelines, or – until very recently - national
European standards published e.g. by BSI (2000) or DIN (2008). These documents contain
limiting values for D/t – see Fig. 6-, below which the CHS can be considered to be (at least)
“semi-compact”, i.e. the yield stress/ elastic capacity can be reached without local buckling.

Figure 6: D/t limits for full (yield) capacity – limit between semi-compact and slender sections.

8
Table 3: D/t limits for non-compact sections acc. to different design standards and different European steel grades.
Steel ² EN BS 5950:2000 DIN 18800-1 AISC (2010) EN 1993-1-6
Grade 1993-1-1
[fy in =235/fy compr. bending compr. bending compr. bending compr. bending
1 2
MPa] (fy in
MPa)
S235 1.00 90.00 94.00 164.00 70.00 90.00 98.00 277.00 43.00 97.00
S355 0.66 59.58 62.23 108.56 46.34 59.58 64.87 183.37 28.46 64.21
S460 0.51 45.98 48.02 83.78 35.76 45.98 50.07 141.51 21.97 49.55
S550 0.43 38.45 40.16 70.07 29.91 38.45 41.87 118.35 18.37 41.45
S690 0.34 30.65 32.01 55.86 23.84 30.65 33.38 94.34 14.64 33.04
S770 0.31 27.47 28.69 50.05 21.36 27.47 29.91 84.54 13.12 29.60
S890 0.26 23.76 24.82 43.30 18.48 23.76 25.88 73.14 11.35 25.61
1. For “medium-length” cylinders in compression, Cx=1,0; 2. For bending: Cx=1,0, x,0=0.3

Whenever necessary (e.g. in the case of the AISC limits), the D/t ratios in Fig. 6 were re-
calculated on the basis of a value of the Young’s Modulus E equal to 210 GPa, as specified in the
Eurocode. As can be seen in the figure, the ratio 2=fy/235 between the standardized yield stress
(in MPa) of the common mild steel in Europe (S235) and of the used material enters the
definition of the limiting value of D/t. This is explicable in light of the definition of slenderness
used in Eq. 1 and the formula for crit in Eq. 6: with a given “plateau” value of the buckling
knockdown factor curve (  x,0 ), the corresponding D/t (~2 r /t) value can be calculated by solving
the expression at the bottom of Fig. 6 for r/t ~ (D/2)/t.
Contrary to the case of mild steel CHS, for HS steel the resulting limit values of D/t are rather
low, see Table 3. This means that many more sections will fall into the slender category
according to these common rules. Furthermore, the HS steels on the European market typically
feature a stress-strain (-) relationship as sketched on the top left corner of Fig. 6, with
comparatively little “over-strength” – in terms of yield stress- compared to the specified
minimum value, low strain hardening (difference between tensile strength and yield stress), and
relatively low elongation at necking u. This means that the background to the D/t limits
themselves might be called into question when applied to (very) high-strength steel grades.

Figure 7: Background tests for the buckling knockdown factor x in EN 1993-1-1 (a), Plantema diagram as
background for the D/t limits for semi-compact CHS in international building codes (b)

9
The known background to the limits in Fig. 6 and – equivalently – to the “plateau” value of the
buckling reduction factor x in EN 1993-1-6 is shown in Fig. 7. Thereby, Fig. 7a is taken from
the commentary to DIN 18800 (Lindner et al. 1994) and shows an extensive collection of 1200
international buckling tests for meridional compression of metallic cylinders; these tests are
compared to the 2 curve of DIN 18800-4, which is quite similar to the current x,N curve for
quality class “B” in EN 1993-1-6 (compare Fig. 4a). The huge scatter – and work undertaken to
reduce it – is commented upon in the mentioned reference. Nevertheless, the plateau value of
 x,0 =0.2 is stated to be appropriate. It shall be noted that some higher-strength steel grades are
included in the plot, as well as some non-ferrous metallic alloys.
The original source / function that serves as justification for the limit values of ca. 90-100 D/t ²
mentioned in most steel design codes is shown in Fig. 7b: the shown relationships were first
given in the work by Plantema (1946) and are based on a series of original tests on mild steel
tubes.

3.2 Buckling knockdown values in various standards

Once a studied (HS steel) CHS is determined to fall into the “slender” cross-sectional range, it
becomes necessary to account for local buckling in the design formulations. Eurocode 3 (general
part 1-1) contemplates this case only in passing, as it refers the user to EN 1993-1-6 and the shell
buckling rules therein; it shall be noted that this is quite impractical and often not fully consistent
with EC3 part 1-1, as the latter code makes use of “effective cross-sections” (effective area Aeff
and section modulus Weff, reduced from the gross area A and the – elastic - section modulus
Welast) when dealing with slender (“class 4” in EC3) sections, yet EN 1993-1-6 contains no rules
for the determination of effective sections.
Other design codes contain more easily applicable rules for either buckling knockdown factors or
effective areas: For the studied cases of isolated compression and bending, both can easily be
rewritten in terms of the reduction factor x. For example, in BS5950-1 (BSI 2000), the
following two expressions are given (rewritten here to be based on S235 steel instead of the
original S275):
0.5
A eff  94 2 
 x,N    (7)
A D / t 
0.25
Weff  164 2 
 x,M    (8)
Welast  D / t 
Similarly, the AISC (2010) specification contains the following rules (in section E and F):
0.038  E 2 34 2
 x,N Q      2 / 3 (9)
f y  (D / t) 3 D / t

4  t 
 plast =  1   when D/t  62.5   2 (10a)
  D

M n,AISC  0.021 E 
 x,M = 1   when D/t  277   2 (10b)
f y  Welast  fy  D / t 
 0.33  E 
  when D/t  277   2 (10c)
 fy  D / t 

10
While section 3 was solely concerned with “local buckling”, it is obvious that CHS elements are
usually (also) globally slender and thus potentially susceptible to flexural column buckling. Very
little evidence has been gathered so far on the interactive local/global buckling of HS steel
cylindrical elements; an extrapolation from findings developed for other steel grades or structural
types is thus not generally recommendable.
In order to overcome these problems, specific work packages dedicated to (HS steel) CHS
sections were included in two on-going or targeted research projects of the authors: the CIDECT-
funded project “Hollopoc” on the “Overall Interaction Concept” (OIC) design of hollow sections
(carried out primarily at HES Fribourg - University of Applied Sciences of Western
Switzerland), and the follow-up “Hollosstab” project specifically dedicated to HS steel (with
additional research partners from IST Lisbon and Imperial College London, as well as from the
industry). Both projects are part of the STSS (2012) research framework, initiated by the authors.
The “Overall Interaction Concept” is explained in a second paper in these same SSRC
conference proceedings (Nseir et al. 2014) and will thus not be discussed in detail here. Briefly,
as the name implies, and Fig. 10a shows, the method again has many analogies with the
EN 1993-1-6 “Overall Method”, as well as with DSM. The focus of the method, however, is
globally and locally slender beam-columns, with combined bending and compression; isolated
load cases (only N, only M) and buckling modes (only local, only global) are included in the
general procedure as special cases.
The “Hollosstab” project, which specifically focuses on HS steel, is currently at the
preparation/initial test phase. Preliminary results for HS steel CHS, developed by means of a
numerical (GMNIA) study, are shown in the following. They give an indication of the general
tendencies to be expected from a more thorough experimental validation, and allow one to
formulate some first proposal formats for design equations.

Figure 11: Methodology and assumptions for the preliminary numerical studies

13
4.2 Preliminary numerical studies – methodology of the GMNIA calculations.
The methodology employed in the preliminary numerical studies of this paper is summarized in
Fig. 11. The following assumptions were made:
i. The shape of the lowest local buckling eigenmode, as shown by an LBA analysis, is
considered to be the most detrimental local imperfection shape for the studied CHS. This
is generally a sufficiently accurate assumption only for cylinders of intermediate local
slenderness, as other patterns may be more critical in the case of either more thick- or
thin-walled elements. The amplitude of the imperfection was chosen to have a value
proportional to the diameter of the shell, in accordance with the tolerance specifications
in EN 10210 and EN 10219.
ii. The global imperfection – relevant for the study of beam-column behavior – was applied
with an amplitude of e0,glob=L/1000. This corresponds to the common assumptions made
during the development of the current EC3 beam-column design rules.
iii. Residual stresses were modelled with the amplitudes shown in the figure. The minimum,
nominal material properties (yield stress fy, elongation at necking u=15 = 15 fy/E,
fu=1.05 fy) were used for HS steel.
iv. For both short and longer elements, a boundary condition corresponding to “thick plates”
and penetration welds (“rotationally fixed shell edges at ends”) was considered; for the
purposes of global buckling, the beam-columns had hinged ends.
v. The applied reference load was equal to the maximum plastic capacity R pl – calculated
omitting any local buckling. In the case of isolated compression, this simply means that
A*fy was applied. In the case of bending, the plastic section modulus Wpl was used and a
moment of M= Wpl * fy was applied. Only a selection of results is discussed in the
following due to space limitations; these results are representative of the general findings.

4.3 First results and proposals

A first series of calculation results is reported in Table 4. A number of HS steel sections with
“common”, small to medium diameters and steel grades up to S770 (fy=770 MPa) were loaded in
axial / meridional compression. All values of D/t exceeded the limit value of 90 ² and the cross-
sections were thus “slender” according to the general part 1-1 of Eurocode 3. The length of the
sections was set to L=6,67 * D / , which leads to globally stocky members with  glob~0.2.
Table 4: Numerical (GMNIA) results for local buckling (globally stocky CHS)-
compression, several fabricated dimensions
CHS Steel Grade L D/t Rcr
 GMNIA
Dimensions
 2

D x t [mm] (fy=ReH,min) [mm] [-] [-] [-]

273 x 8,0 S770 – 1006.0 111.8 8.753 0.3380 1.004
(fy=770MPa)
177,8 x 5,0 S770 655.2 116.5 8.4324 0.3444 1.004
355,6 x 8,0 S770 1310.3 145.6 7.0000 0.3780 0.989
200 x 4,0 S500 914.5 106.4 9.3264 0.3274 0.9978
200 x 3,0 S500 914.5 141.8 7.1630 0.3736 0.9890
273 x 4,0 S700 1055.1 203.3 5.0607 0.4445 0.9770
244,5 x 5,0 S700 944.9 145.7 6.9812 0.3785 0.9910

14
The obtained values of Rcr,  and GMNIA are shown in the table. It is evident, from these first
calculations, that the studied sections had a rather small tendency to buckle in a local mode: all
reduction factors are very close to 1.0.

In a second step, fictitious sections with D=300 and progressively reduced wall thickness, all of
steel grade S690, were looked at in order to obtain GMNIA results with a more severe effect of
local buckling. These results are summarized in Table 5

Table 5: Numerical (GMNIA) results for local buckling (globally stocky CHS)-
compression, D=300mm, variable thickness
CHS Steel Grade L D/t Rcr
 GMNIA
Dimensions 2

D x t [mm] (fy=ReH,min) [mm] [-] [-] [-]
300 x 10,0 88.1 10.976 0.302 1.002
300 x 8,0 110.1 8.910 0.335 1.000
300 x 6,0 146.8 6.925 0.380 0.989
300 x 4,0 220.2 4.726 0.460 0.956
S690 1167.8
300 x 3,0 (fy=690MPa) 293.6 3.594 0.528 0.925
300 x 2,0 440.4 2.442 0.640 0.881
300 x 1,75 503.3 2.150 0.682 0.863
300 x 1,50 587.2 1.856 0.734 0.813

It was consequently attempted to describe these results with an analytical formula, based on the
EN 1993-1-6 buckling reduction factors, but transferred into a format that is more suitable for the
planned “OIC” type of representation.
For this purpose, as a first proposal, the well-known “Ayrton-Perry” format was chosen, as this
format can also very advantageously be used for the representation of (beam-)column reduction
factors. The following expressions were thus calibrated:

1
 x,loc   1.0 (11)
2
 loc
 loc   loc 2 
1
  loc 
2

with  loc
 
 0.5  1   0  loc   loc,0 

1 
 (12)
 

0  0.5 ; 
1  0.8  1  2  w k / t 
1.5
 ; w k / t 
1
Q
r
t
(13)

The plateau value  loc,0 is at first deliberately left undefined, but will typically take values
between 0.2 and 0.4. Note that wk/t is taken directly from Eq. 3b, while 1 corresponds to a
slightly modified version of Eq. 3a. The index “loc” is introduced here to distinguish the
reduction factor for local buckling and its coefficients from the corresponding factors for global
(column) buckling.

15
 Figure 12: Approximation of the EN 1993-1-6 local buckling rules by an alternative, more compact formulation.

The results of this approximation are shown first for a purely analytical evaluation of the
slenderness  loc =f(D/t) and of the EC3-1-6 reduction factor x=loc, using Eq. 6 and Eq. 1,
Cx=0.6, and  loc,0=0.2, see Fig. 12. The figure clearly illustrates the accuracy of the
approximation.

In a subsequent step, the results of Table 5 are compared with the EC3-1-6 results and the above
approximation calculated in two different ways:
i. The EC3-1-6 results were calculated via crit obtained using the approximation for Cx
given in the code, i.e. the red line in Fig. 5, with the appropriate factors for the given
“fixed shell edge” boundary condition, length and D/t ratios. Furthermore, Eq. 2 and the
standard coefficients in Table 1 and 2 for , ,  and  0 (all for meridional compression)
were used. This leads to the dark continuous line in Fig. 13.
ii. The same procedure for the calculation of crit was carried out for the above
approximation of the EC3-1-6 local buckling rules, using Eq. 11 to 13 with  loc,0=0.2.
This is represented by the red dashed line in Fig. 13, again showing the accuracy of the
approximation (“Approxim. v1”) in describing the EC3 curves – yet not quite in
describing the GMNIA results.
iii. Finally, the approximation of Eq. 11 to 13 was evaluated directly on the basis of  loc
from the numerical LBA analysis, i.e. without the need for explicitly calculating crit
from Eq. 6. Additionally, and more importantly, the plateau value of  loc,0 was raised to
 loc,0=0.4. This leads to the “Approxim. v2” points in Fig. 13, which match the GMNIA
results quite closely.
It shall be noted that a value of  loc,0=0.4 more closely matches the D/t=90 ² value of EN 1993-
1-1 (and the similar limit/”plateau” values of D/t in other codes).

16
 Figure 13: Local buckling of HS steel CHS in compression; GMNIA vs. analytical predictions.

Finally, some first results concerning the behavior of CHS beam-columns are presented in the
following. For CHS, the main problems to be dealt with are: i. the interaction between local and
global buckling, and ii. the effect of the bending moment in a representation of the beam-column
strength in terms of the “Overall Interaction Concept”, i.e. on the basis of a combined, overall
slenderness equivalent to the one given by Eq. 1.
Both aspects are separately dealt with in the two diagrams of Fig. 14:
i. Fig. 14a treats the interaction between local and global buckling of a HS steel CHS with
D/t=273/5=54.6. With steel grade S690, this section falls clearly outside of the D/t lmit
value of 90 ²=30.7, thus making it necessary to consider local/global interactive
buckling. This is done using the following equations 14 to 16, which apply the
conventional EC3 approach (comparable to the “Q” factor approach of AISC).

 
1
glob,gr   x,loc  MIN  ;1.0 (14)
 2 
  glob   glob   glob
2

with   
g lob  0.5  1  glob  globloc  0.2   globloc
2
 (15)

 x,loc  A  f y
 globloc    glob,gr   x,loc (16)
2 EI / Lcr 2

In the above equations, the glob,gr is the global reduction factor, including effects of local
buckling, defined as the reduction of the gross cross-sectional resistance A*fy. x,loc is the
purely local reduction factor defined above.

17
Figure 14: Buckling strength of beam-columns using the “Overall Interaction Concept”; locally and globally slender
column (a); beam-column loaded by eccentric axial forces (b).

Two proposals are made in Fig. 14a order to describe the GMNIA results: both make use
of the “exact” numerical slenderness retrieved from the LBA analysis, but the lower of
the two proposals uses  loc,0=0.2, while the higher one uses  loc,0=0.4. The increased
accuracy of the latter can again be appreciated. These proposals are compared to the
existing EC3 rule, combined with the simple effective area formula of BS 5950-1, Eq. 7.
The resulting curve is plotted as a blue dashed line in Fig. 14a.

ii. Fig. 14b deals with the behavior of relatively stocky CHS beam-columns, loaded by
eccentric axial forces at both ends, resulting in compression plus a uniform first-order
bending moment diagram. The GMNIA results are shown by the grey line with circle
marks. A proposal was made in (Taras 2010; Taras & Greiner 2010) which leads to a
very accurate description of the beam-column’s buckling strength in terms of the
“overall” slenderness and reduction factor  ov and ov. While the details cannot be
included in this paper, it shall be mentioned that the key feature of the formulation is the
inclusion of the “load eccentricity”, normalized according to Eq. 17 by the plastic
section core width Mpl/Npl, into the Ayrton-Perry imperfection term glob globloc  0.2 
of Eq. 15.

0 
 M / M pl   m0 (17)
 N / N pl  n 0
Note that no local buckling is yet included in this proposal. However, the results for
stockier beam-columns and for locally slender columns are encouraging for the feasibility
of a combination of the two solutions.

18
5. Summary, Conclusions and Outlook
In the first part of this paper, the Eurocode 3 – Part 1-6 “Overall Method” for the design of steel
shell structures against buckling was described in this paper and compared to other methods that
share the same basic concepts, namely the use of numerical analysis techniques for the precise
calculation of the elastic bifurcation load of the studied system and the application of a global
buckling reduction factor to the (plastic) collapse load of the system. The analogies and
differences between this design method for thin-walled shell structures and the Direct Strength
Method (DSM) for thin-walled, cold-formed “flat” cross-sections were discussed.
The second part of the paper discussed the need for the consideration of shell buckling
resistances in the design of only moderately thin-walled circular hollow sections (CHS) made of
high-strength steel with yield stress values of up to and beyond S770. These sections may often
fall below the common slenderness limits D/t between semi-compact and slender cross-sections.
It is therefore necessary to account for local buckling, as well as global buckling whenever the
studied member is also susceptible to global, flexural column buckling.
Two on-going or starting research projects were presented, which deal specifically with the
resistance of slender, high-strength CHS, making use of the “Overall Interaction Concept”, a
concept for the design of slender beam-columns that is comparable to the DSM. Initial results
were shown, based on preliminary numerical studies, and possible formats for design equations
were presented. The mentioned research projects will seek to answer the following open
questions concerning the design of slender, high-strength CHS, namely:
i. What is the influence of the characteristic, low-ductility stress-strain curve of HS
steels on the applicable slenderness limits for the on-set of elasto-plastic buckling
(semi-compact to slender cross-section transition) of CHS?
ii. In the semi-compact and compact range: is enough ductility present to allow for
(partial) plastic design of HS steel CHS members?
iii. Can the concept of classification of cross-sections – and the corresponding non-
continuous strength curves for members of increasing local slenderness – be replaced
by a continuous strength curve, as function of the “Overall Slenderness”, i.e. using
the “Overall Interaction Concept”?
iv. Can the design of locally and globally slender HS steel hollow-section beam-columns
be accommodated within the OIC? Can variable bending moment diagrams,
respectively stress gradients in longitudinal direction, be incorporated into easy-to-use
design rules?
v. Are the local imperfection amplitudes which are currently explicitly or implicitly
contained in existing buckling design rules for cylindrical structures representative of
the scatter-band of actual geometric shape deviations in CHS fabricated to EN 10210,
EN 10219 or comparable other international standards, or can reduced imperfections
be used in design?
vi. Do the lower strength reserves on the yield stress (actual vs. nominal, minimum
value) to which HS steel grades are produced have a detrimental impact on the safety
factors to be used in design?
Acknowledgments
The authors gratefully acknowledge the contributions to and interest in these on-going research
activities by the involved industrial partners, in particular Dr. G. Iglesias at Grupo Condesa S.A.,
Mr. A. Giacobbe and Mr. E. Anelli at Tenaris Dalmine Spa, and Dr. A. Seyr at Voestalpine
Krems GmbH.

19
References
AISC (2010) - ANSI/AISC 360-10, “Specification for Structural Steel Buildings”, American Institute of Steel
Construction, Chicago (IL).
BSI (2000), BS 5950-1, “Structural use of steelwork in buildingCode of practice for design. Rolled and welded
sections”, British Standards Institute, London.
CEN (2005) - EN 1993-1-1, “Eurocode 3. Design of steel structures – Part 1-1: General rules and rules for
buildings”, CEN - European Committee for Standardization, Brussels
CEN (2006a), EN 10210 “Hot finished structural hollow sections of non-alloyed and fine grain”, CEN – European
Committee for Standardization, Brussels.
CEN (2006b), EN 10219 “Cold formed welded structural hollow sections of non-alloy and fine grain steels”, CEN –
European Committee for Standardization, Brussels.
CEN (2010), EN 1993-1-6 – “Eurocode 3: Design of steel structures – Part 1-6: Strength and stability of shell
structures”, CEN - European Committee for Standardization, Brussels
CIDECT (2010), “Hollow Sections in Structural Applications”, J. Wardenier, J.A. Packer, X.-L. Zhao and G.J. van
der Vegte Ed., Comitè International pour le Développement et l’Etude de la COnstruction Tubulaire – CIDECT,
Geneva.
DIN (2008), DIN 18800-1, “Stahlbauten – Teil 1: Bemessung und Konstruktion”; DIN Institute, Berlin.
Flügge, W. (1973), “Stresses in Shells”, 2nd Ed., Springer New York Heidelberg Berlin.
Lindner, J., Scheer, J., Schmidt, H., Ed. (1994), „Stahlbauten. Erläuterungen zu DIN 18800 Teil 1 bis Teil 4
[Commentary to DIN 18800 part 1 to 4, in German] ”, Beuth, Berlin.
Nseir, J., Saloumi, E., Hayeck, M., Taras, A., Boissonnade, N. (2014), “An Alternative Approach for the Prediction
of Hollow Structural Shapes Cross-sectional Resistance: the Overall Interaction Concept”, Proc. SSRC Annual
Stab. Conf. , Toronto, Canada, March 25-28, 2014.
Plantema, F. J., "Collapsing Stress of Circular Cylinders and Round Tubes," National Luchtraat Laboratorium,
Report S .280, Amsterdam, Netherlands, 1946.
Rotter, J. M. (2011), “Shell buckling design and assessment and the LBA-MNA methodology”. Stahlbau, 80, pp.
791–803, Ernst&Sohn / Wiley, Berlin.
Rotter, J. M., Schmidt, H. Ed. (2008), “European Recommendations for Steel Construction: Buckling of Shells, 5th
edition”, European Convention for Constructional Steelwork, Brussels.
S.T.S.S. (2012). “Research project STSS – Simple Tools Sell Steel” (http://www.ims.org/2012/02/stss-simple-tools-
sell-steel/).
Schafer, B.W., (2008) “Review: The Direct Strength Method of cold-formed steel member design”, Journal of
Constructional Steel Research, Vol. 64, Issue 7-8, pp. 766-778, Elsevier.
Semi-Comp (2007), “Plastic member capacity of semi-compact steel sections – a more economic design (Semi-
Comp)” . Final report (01/01/06 – 30/06/07) – RFCS – Steel RTD (Contract RFS-CR-04044).
Taras, A. (2010), “Contribution to the Development of Consistent Stability Design Rules for Steel Structures”, PhD
Thesis, Graz University of Technology, Austria.
Taras, A., Greiner, R. (2010), “Analytical Derivation of a Generalized Slenderness Formula for In-Plane Beam-
Column Design and Comparison with Interaction-Concept Formulae”, Proc. of Stability and Ductility of Steel
Structures – SDSS 2010, Rio de Janeiro, Brazil, September 8 - 10, 2010.
Timoshenko, S. P. (1934), Theory of Elasticity, 1st Ed., McGraw-Hill Book Company.

20